Research platforms have made surveys easier to distribute, but easier distribution has not solved the harder problem: getting people to stay engaged long enough to give thoughtful answers. 60% of respondents do not want to take a survey lasting more than 10 minutes, while 87% do not want one lasting more than 20 minutes. That creates a direct incentive to ask fewer, better questions.
For Nathan Koengeter, the lesson is familiar from a different setting. His background includes in-person canvassing, political polling, qualitative and quantitative market research, academic research and analytics. That combination has kept his work close to a basic truth about surveys: before a response becomes a row in a dataset, a person has to understand the question, decide what it means and find an answer that actually fits.
The Questionnaire Is Part of the Measurement
Survey wording is not just packaging. AAPOR recommends that questions be short, simple, specific, free of bias and limited to one concept at a time. It also advises researchers to use response options that cover reasonable answers and to pretest questionnaires with people similar to the eventual respondents.
That guidance aligns closely with Nathan Koengeter’s field experience. In-person research makes confusion visible. A researcher can hear hesitation, notice when a respondent interprets a term differently than intended or recognize that the available choices do not capture what the person is trying to say. In an online survey, those same problems can disappear into a clean-looking spreadsheet.
Nathan Koengeter has described canvassing as a form of qualitative research because each interaction provides more than a binary response. It exposes the vocabulary people use, the distinctions they make and the context surrounding an opinion. For Nathan Koengeter, those observations are useful raw material for stronger quantitative research later.
Write for Respondents, Not for the Research Team
Researchers naturally become fluent in the language of the topic they are studying. Customers and respondents do not have that obligation. A phrase that feels precise inside a company can be vague, technical or loaded to somebody outside it.
Nathan Koengeter approaches this as an input-quality problem. If the respondent has to translate the question before answering it, the research team has introduced avoidable uncertainty before analysis even begins. Better questions use language that is familiar to the audience without sacrificing precision.
This is also why Nathan Koengeter pays attention to the language that appears in interviews and field conversations. If customers repeatedly describe a problem one way while the organization describes it another, the gap itself may be important. Survey design should begin with the language of the people being studied, not simply the vocabulary already used in a boardroom.
One Question Should Measure One Thing
A common survey mistake is asking two things at once. A question such as whether a service was “fast and helpful” gives one response box to a customer who may believe it was fast but unhelpful. The resulting number looks precise while hiding an unresolved distinction.
AAPOR specifically advises researchers to ask about one concept at a time. Nathan Koengeter sees the same principle later in the data pipeline: ambiguous inputs create ambiguous outputs. A reporting system cannot reconstruct meaning that the questionnaire failed to capture.
Different Research Modes Can Produce Different Answers
How a question is delivered can matter as much as how it is written. Ipsos notes that different modes can create both mode effects and selection effects because respondents may interpret questions differently in interviewer-administered and self-completed settings, while different recruitment channels can attract different groups. A result can therefore move because the research method changed, not because the underlying opinion changed.
That is especially relevant to Nathan Koengeter because his experience spans face-to-face fieldwork, polling and structured market research. Koengeter does not assume that an answer collected at a doorstep, by phone and through an online questionnaire is automatically interchangeable. The method is part of the evidence.
Leave Room for the Answer You Did Not Expect
Closed-ended questions are efficient because they make comparison easy. They also force the researcher to decide in advance which answers exist. That can be useful after the important categories are understood and dangerous before they are.
Nathan Koengeter’s broader research positioning combines qualitative and quantitative methods for this reason. Interviews, open-ended responses and field research can uncover motivations or distinctions that were not obvious at the beginning. Quantitative research can then test how widely those patterns hold.
The practical case for restraint is visible in survey design data as well. 26,000 SurveyMonkey Audience surveys found a median length of 12 questions; 75% had 21 questions or fewer. More questions are not automatically more insight.
Better Surveys Begin Before the First Response
The most expensive survey error is often not a bad calculation. It is asking the wrong question clearly and collecting thousands of precise answers to it.
Nathan Koengeter’s career has moved from direct field interactions to structured market research and large-scale analytics, but the logic has remained consistent. Reliable decisions depend on reliable inputs. In survey research, Koengeter argues that reliability begins with questions real people can understand, answer choices that reflect reality and a design that respects the respondent’s time.
